Language in Context: MEG Evidence for Modality-General and -Specific Responses to Reference Resolution

Abstract

Successful language comprehension critically depends on our ability to link linguistic expressions to the entities they refer to. Without reference resolution, newly encountered language cannot be related to previously acquired knowledge. The human experience includes many different types of referents, some visual, some auditory, some very abstract. Does the neural basis of reference resolution depend on the nature of the referents, or do our brains use a modality-general mechanism for linking meanings to referents? Here we report evidence for both. Using magnetoencephalography (MEG), we varied both the modality of referents, which consisted either of visual or auditory objects, and the point at which reference resolution was possible within sentences. Source-localized MEG responses revealed brain activity associated with reference resolution that was independent of the modality of the referents, localized to the medial parietal lobe and starting ∼415 ms after the onset of reference resolving words. A modality-specific response to reference resolution in auditory domains was also found, in the vicinity of auditory cortex. Our results suggest that referential language processing cannot be reduced to processing in classical language regions and representations of the referential domain in modality-specific neural systems. Instead, our results suggest that reference resolution engages medial parietal cortex, which supports a mechanism for referential processing regardless of the content modality.

Cortical processing of reference in language revealed by computational models

“…Both our fMRI and MEG results showed significant LMTG activity, consistent with previous literature 14,15,20 on pronoun resolution. Our MEG results showed additional activity in the left medial parietal lobe, which also replicated previous MEG results 17,18 . We extracted the LMTG and the RMTG clusters from the fMRI results as the functional regions of interests (fROIs) for further representational similarity analyses on the fMRI data.…”

“…The left STG, MTG and IFG have been shown to elicit increased activation with increased linear distance between pronouns and their antecedents 14,15 , and the left AG and PC have been implicated in backward anaphora processing 16 . The MEG literature has mainly localized reference resolution to the medial parietal lobe 17,18 . However, no consensus has been reached on the exact location of third person pronouns, which are the focus of the current study.…”

“…Pronouns are ubiquitous in human language, yet despite decades of research on pronoun resolution in linguistics and cognitive science, the neural basis for pronoun processing remains largely uncharacterised, and it is absent from recent neurobiological models of language comprehension (e.g., Friederici, 2011; Hagoort & Indefrey, 2014). Prior neuroimaging studies on referential processing have reported a number of regions, including the medial parietal lobe (van Berkum et al, 2007; Brodbeck & Pylkkänen, 2017; Brodbeck et al, 2016), the lateral parietal region (van Berkum et al, 2007), the inferior frontal region (Hammer et al, 2007; Matchin et al, 2014; Miceli et al, 1991) and the temporal regions (Hammer et al, 2007, 2011; Miceli et al, 1991), yet they involve different experimental manipulations such as contrasting referentially ambiguous pronouns with non-ambiguous pronouns (van Berkum et al, 2007), contrasting gender-incongruent with gender-congruent pronouns (Hammer et al, 2007). It is therefore unclear whether these manipulations tapped the same cognitive processes.…”

“…Given that the four models focus on different aspects of pronoun resolution, it may well be the case that they are distinctively correlated with different brain regions in a broad network for pronoun processing. To search for regions associated with each model, we conducted a searchlight RSA within a mask covering regions that have been previously reported for pronoun processing, including the STG, MTG, IFG, AG and PC 14,15,16,17,22 (see "Searchlight RSA: fMRI data" in Methods for details).…”

Neural basis of basic composition: what we have learned from the red–boat studies and their extensions

“…The more straightforward part of the project would be assessing whether LIFG effects might emerge in MEG when a grammaticality judgement task is used. The two additional regions revealed by our work in contrast with the fMRI language map are the vmPFC, as discussed above, and the medial parietal cortex, revealed in a series of studies aimed at identifying neural correlates of reference resolution [68,69]. Importantly, both of these regions also belong to the default network, comprising a set of regions activated when subjects are left to just think to themselves [73].…”